We are using version 5.5.
I ran follwoing reports twice with the same exact configration and nothings change but still I get big difference in Response Time.
<font class="small">Code:</font><hr /><pre>
Run1 Run2 Percentage difference
Report1 3.1106 2.5099 -23.93%
Report2 3.3365 2.7281 -22.30%
Report3 3.9135 3.0021 -30.36%
Report4 3.8875 3.1449 -23.61%
Report5 3.6530 2.7165 -34.47%
Report6 3.6169 2.9084 -24.36%
Does a 3rd run produce results similar to the 2nd?
Just curious if there is a lot of caching or spin-up that accounts for faster response times in subsequent runs after the 1st.
I forgot to mention one thing is that we bounce the db and app server after each run.
We never did the 3rd run.
I would try more iterations (perhaps without bouncing in between.. you want a warmed up environment). See if 3rd, 4th, etc runs give different results.
My first thought would be server side caching. Are you sure that this is being cleared out? It isn't completely apples-to-apples, but try running different reports for each iteration.
If you can reproduce this, then there is something different for the first run (such as 1 of the servers/apps/services is "warmed up," etc.).
A problem is a difference between what is perceived and what is desired, that
we want to reduce (Dewey 1933)